Robust reliability optimization of mechanical components using non-probabilistic interval model

Abstract

Mechanical components design is subjected to uncertainties in material and geometrical properties, loads and other variables. For reliability optimization design with uncertain parameters, based on the non-probabilistic reliability theory, robust design and optimal design method, the uncertain parameters of mechanical components are expressed by non-probabilistic interval variables, and a non-probabilistic measure and procedure for robust reliability computation is presented. Compared with the conventional probabilistic reliability optimization approach, the proposed method does not require a presumed probability distribution of the uncertain parameters and only the bounds or ranges of their variations are required. The optimal design for non-probabilistic robust reliability is formulated as a two level optimization problem, in which the first level minimizes the original robust optimal objective with the constraints of non-probabilistic reliability index, and the secondary level is used to identify the reliability index. The purpose of it is to get a tradeoff between the design objective and the robustness to uncertainties and satisfy the requirements for reliability. For instance, the robust reliability optimization of unidirectional wedge-typed overrunning clutch is proposed, and the results show that the proposed method is useful for mechanical components design and quality improvement.